Recherche avancée

Sur d’autres sites (3518)

• Trouble with Cross-Origin Headers : Stripe and ffmpeg Integration in Next.js

  7 février 2024, par Duke

  I'm using this excellent template from vercel as a base for my project :

  


  https://vercel.com/templates/next.js/subscription-starter


  


  I've used it for many projects and the Stripe integration works flawlessly. However for a new project I needed to incorporate ffmpeg which relies on SharedArrayBuffer :

  


  https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/SharedArrayBuffer


  


  So I need to set the appropriate HTTP response header which i've set in my next.config.js file as below :

  


        {
        source: '/(.*)',
        headers: [
          {
            key: 'Cross-Origin-Embedder-Policy',
            value: 'credentialless'
            // value: 'require-corp'
          },
          {
            key: 'Cross-Origin-Opener-Policy',
            value: 'same-origin-allow-popups'
          }
        ]
      }


  


  Now the issue is that settings these breaks Stripe. It blocks the request :

  


  https://js.stripe.com/v3/m-outer-3437aaddcdf6922d623e172c2d6f9278.html


  


  Which appearently loads more resources that are needed for the checkout session to function. The customer portal works but checkout doesn't. I've literally tired every possible permutation of headers with different values but in all cases either Stripe works but ffmepg doesn't (cause SharedArrayBuffer doesn't work) or vice versa. It's been a week and I've tired everything and can't find anything online that works. Not even ChatGPT can help which ironcelly asks me to contact Stripe customer support. I would highly appreciate if anyone could even point me in the right direction. I Thank you in advance good Sir/Maam.

  


  Note that Checkout works fine with just :

  


  key: 'Cross-Origin-Opener-Policy',
value: 'same-origin'


  


  Just Cross-Origin-Embedder-Policy breaks it

  


• Using FFmpeg to detect alpha channel in PR4444 file [closed]

  8 décembre 2023, par Justin Myers

  I wanted to comment on a thread that already exists on this topic in general, but I just signed up to SO and have not earned my contribution credits yet ! If anyone has advice on how to better handle that in the future, I am all ears !!

  


  None-the-less, here is that thread :

  


  A good way to detect alpha channel in a video using ffmpeg/ffprobe

  


  I tried both the answers listed in that post to detect the presence of an alpha channel in a PR4444 video file.

  


  For Gyan's answer, I get the following pixel format from any PR4444 file (regardless if encoded with alpha or no alpha), when running part 1 :

  


  yuva444p12le

  


  As you will see in Gyan's notes, even if a FFprobe returns an "a" in the return string (for the call for pixel format), that does not equate to the presence of an alpha channel. Hence their note for the subsequent (part 2) call...

  


  I then plugged that into the part 2 of Gyan's answer, but that produced an empty from FFmpeg, and a message providing valid arguments to pair with grep. It seems as though -oP is not a valid arg for grep ? Because it wasn't listed in the returned list of options. This is the template I used (pulled verbatim from Gyan's posted solution) :

  


  ffprobe -v 0 -show_entries pixel_format=name:flags=alpha -of compact=p=0 | grep "$PIX_FMT|" | grep -oP "(?<=alpha=)\d"

  &#xA;

  Where $PIX_FMT is to be replaced with yuva444p121e (per Gyan's instructions).

  &#xA;

  I then tried Benji's solution for part 2, but that just returns the following :

  &#xA;

  pix_fmt=yuva444p12le

  &#xA;

  Of course, this isn't any more useful than part 1...

  &#xA;

  I suspect Gyan is on the right track as he mentions part 2 should produce a boolean result for the presence of an alpha channel. However, either I am misunderstanding the syntax of his template, or something has changed in FFprobe since. There is so limited information out there for this specific task. Hoping someone with more experience and knowledge can help shed some light for me ??

  &#xA;

• ffmpeg output seeking not accurate even after -i [closed]

  3 septembre 2023, par hl037_

  In the documentation of ffmpeg, it is stated that if -ss is passed after -i, then it is frame accurate (Other questions are for -ss before -i).

  &#xA;

  Here is my command :

  &#xA;

  ffmpeg -i vid.mkv -ss 53 -codec copy test.mkv -y 


  &#xA;

  However, it is still not accurate, and seems to jump to the nearest keyframe. Am i missing something ?

  &#xA;

  Note : the problem only occurs when the video steam is copied. If reencoded or discarded, it works as expected

  &#xA;

  Note2 (As I am doing other tries) : There are also strange things with audio :

  &#xA;

  ffmpeg -i vid.mkv -ss 53 -codec copy test1.mkv -y 
ffmpeg -i vid.mkv -ss 54 -codec copy test2.mkv -y 
ffmpeg -i test1.mkv -vn -c:a copy test1a.mkv -y
ffmpeg -i test2.mkv -vn -c:a copy test2a.mkv -y


  &#xA;

  &#xA;
  • test1.mkv is reported to be the right lenght, but the first frame correspond to the frame at 0:50 in original (end is good)

  &#xA;

  • test2.mkv is reported to be the right lenght, but the first frame correspond to the frame at 0:55 in original (end is good)

  &#xA;

  • test1a.mkv is the expected audio (starts at 0:53 of the original)

  &#xA;

  • test2a.mkv is the expected audio (starts at 0:54 of the original)

  &#xA;

  &#xA;